hahaju/aiagents-stock
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

增加龙虎榜批量分析功能，修复一些小bug

Browse Source
This commit is contained in:
oficcejo
2025-10-20 15:40:14 +08:00
parent 1bddf393e0
commit b7a8bb10b3
60 changed files with 3579 additions and 883 deletions
Download Patch File Download Diff File Expand all files Collapse all files
longhubang_ui.py
+595 -13
View File
@@ -102,6 +102,11 @@ def display_longhubang():
def display_analysis_tab():
"""显示分析标签页"""
# 检查是否触发批量分析（不立即删除标志）
if st.session_state.get('longhubang_batch_trigger'):
run_longhubang_batch_analysis()
return
st.subheader("🔍 龙虎榜综合分析")
# 参数设置
@@ -383,6 +388,33 @@ def display_scoring_ranking(result):
use_container_width=True
)
# 一键批量分析功能
st.markdown("---")
col_batch1, col_batch2, col_batch3 = st.columns([2, 1, 1])
with col_batch1:
st.markdown("#### 🚀 批量深度分析")
st.caption("对TOP10股票进行完整的AI团队分析，获取投资评级和关键价位")
with col_batch2:
batch_count = st.selectbox(
"分析数量",
options=[3, 5, 10],
index=0,
help="选择分析前N只股票"
)
with col_batch3:
st.write("") # 占位
if st.button("🚀 开始批量分析", type="primary", use_container_width=True):
# 提取股票代码
stock_codes = top10_df.head(batch_count)['股票代码'].tolist()
# 存储到session_state，触发批量分析
st.session_state.longhubang_batch_codes = stock_codes
st.session_state.longhubang_batch_trigger = True
st.rerun()
st.markdown("---")
# 评分分布图表
@@ -687,32 +719,223 @@ def display_pdf_export_section(result):
def display_history_tab():
"""显示历史报告标签页"""
"""显示历史报告标签页（增强版）"""
st.subheader("📚 历史分析报告")
try:
engine = LonghubangEngine()
reports_df = engine.get_historical_reports(limit=20)
reports_df = engine.get_historical_reports(limit=50)
if reports_df.empty:
st.info("暂无历史报告")
return
st.dataframe(
reports_df,
column_config={
"id": st.column_config.NumberColumn("ID", format="%d"),
"analysis_date": st.column_config.TextColumn("分析时间"),
"data_date_range": st.column_config.TextColumn("数据日期范围"),
"summary": st.column_config.TextColumn("摘要")
},
hide_index=True,
use_container_width=True
)
st.info(f"💾 共有 {len(reports_df)} 条历史报告")
# 显示报告列表
st.markdown("### 📋 报告列表")
# 为每条报告创建展开面板
for idx, row in reports_df.iterrows():
report_id = row['id']
analysis_date = row['analysis_date']
data_date_range = row['data_date_range']
summary = row['summary']
# 创建展开面板
with st.expander(
f"📄 报告 #{report_id} | {analysis_date} | 数据范围: {data_date_range}",
expanded=False
):
# 获取完整报告详情
report_detail = engine.get_report_detail(report_id)
if not report_detail:
st.warning("无法加载报告详情")
continue
# 显示摘要
st.markdown("#### 📝 报告摘要")
st.info(summary)
st.markdown("---")
# 显示推荐股票
recommended_stocks = report_detail.get('recommended_stocks', [])
if recommended_stocks:
st.markdown(f"#### 🎯 推荐股票 ({len(recommended_stocks)}只)")
# 创建DataFrame显示
df_stocks = pd.DataFrame(recommended_stocks)
st.dataframe(
df_stocks,
column_config={
"rank": st.column_config.NumberColumn("排名", format="%d"),
"code": st.column_config.TextColumn("代码"),
"name": st.column_config.TextColumn("名称"),
"net_inflow": st.column_config.NumberColumn("净流入", format="%.2f"),
"reason": st.column_config.TextColumn("推荐理由"),
"confidence": st.column_config.TextColumn("确定性"),
"hold_period": st.column_config.TextColumn("持有周期")
},
hide_index=True,
use_container_width=True
)
st.markdown("---")
# 尝试解析完整分析内容
analysis_content_parsed = report_detail.get('analysis_content_parsed')
if analysis_content_parsed and isinstance(analysis_content_parsed, dict):
# 显示AI分析师团队报告
agents_analysis = analysis_content_parsed.get('agents_analysis', {})
if agents_analysis:
st.markdown("#### 🤖 AI分析师团队报告")
agent_info = {
'youzi': {'title': '🎯 游资行为分析师', 'icon': '🎯'},
'stock': {'title': '📈 个股潜力分析师', 'icon': '📈'},
'theme': {'title': '🔥 题材追踪分析师', 'icon': '🔥'},
'risk': {'title': '⚠️ 风险控制专家', 'icon': '⚠️'},
'chief': {'title': '👔 首席策略师', 'icon': '👔'}
}
for agent_key, info in agent_info.items():
agent_data = agents_analysis.get(agent_key, {})
if agent_data:
with st.expander(f"{info['icon']} {info['title']}", expanded=False):
analysis = agent_data.get('analysis', '暂无分析')
st.markdown(analysis)
st.caption(f"分析时间: {agent_data.get('timestamp', 'N/A')}")
# 显示AI评分排名
scoring_ranking = analysis_content_parsed.get('scoring_ranking', [])
if scoring_ranking:
st.markdown("---")
st.markdown("#### 🏆 AI智能评分排名 (TOP10)")
df_scoring = pd.DataFrame(scoring_ranking[:10])
# 显示完整的评分表格
st.dataframe(
df_scoring,
column_config={
"排名": st.column_config.NumberColumn("排名", format="%d"),
"股票名称": st.column_config.TextColumn("股票名称", width="medium"),
"股票代码": st.column_config.TextColumn("代码", width="small"),
"综合评分": st.column_config.NumberColumn(
"综合评分",
format="%.1f",
help="总分100分"
),
"资金含金量": st.column_config.ProgressColumn(
"资金含金量",
format="%d分",
min_value=0,
max_value=30
),
"净买入额": st.column_config.ProgressColumn(
"净买入额",
format="%d分",
min_value=0,
max_value=25
),
"卖出压力": st.column_config.ProgressColumn(
"卖出压力",
format="%d分",
min_value=0,
max_value=20
),
"机构共振": st.column_config.ProgressColumn(
"机构共振",
format="%d分",
min_value=0,
max_value=15
),
"加分项": st.column_config.ProgressColumn(
"加分项",
format="%d分",
min_value=0,
max_value=10
),
"顶级游资": st.column_config.NumberColumn("顶级游资", format="%d家"),
"买方数": st.column_config.NumberColumn("买方数", format="%d家"),
"机构参与": st.column_config.TextColumn("机构参与"),
"净流入": st.column_config.NumberColumn("净流入(元)", format="%.2f")
},
hide_index=True,
use_container_width=True
)
# 显示评分说明
with st.expander("📖 评分维度说明", expanded=False):
st.markdown("""
**AI智能评分体系 (总分100分)**
- **资金含金量** (0-30分)：顶级游资+10分，知名游资+5分，普通游资+1.5分
- **净买入额** (0-25分)：根据净流入金额大小评分
- **卖出压力** (0-20分)：卖出比例越低得分越高
- **机构共振** (0-15分)：机构+游资共振15分最高
- **加分项** (0-10分)：主力集中度、热门概念、连续上榜等
💡 评分越高，表示该股票受到资金青睐程度越高！
""")
# 显示数据概况
data_info = analysis_content_parsed.get('data_info', {})
if data_info:
st.markdown("---")
st.markdown("#### 📊 数据概况")
col1, col2, col3 = st.columns(3)
with col1:
st.metric("龙虎榜记录", f"{data_info.get('total_records', 0)} 条")
with col2:
st.metric("涉及股票", f"{data_info.get('total_stocks', 0)} 只")
with col3:
st.metric("涉及游资", f"{data_info.get('total_youzi', 0)} 个")
else:
# 如果无法解析，显示原始内容
st.markdown("#### 📄 原始分析内容")
analysis_content = report_detail.get('analysis_content', '')
if analysis_content:
st.text_area("", value=analysis_content[:2000], height=200, disabled=True)
if len(analysis_content) > 2000:
st.caption("(内容过长，仅显示前2000字符)")
# 导出按钮
st.markdown("---")
col_export1, col_export2 = st.columns(2)
with col_export1:
if st.button(f"📥 导出为PDF", key=f"export_pdf_{report_id}"):
st.info("PDF导出功能开发中...")
with col_export2:
if st.button(f"📋 加载到分析页", key=f"load_report_{report_id}"):
# 将历史报告加载到当前分析结果中
if analysis_content_parsed:
# 重建完整的result结构
loaded_result = {
"success": True,
"timestamp": report_detail.get('analysis_date', ''),
"data_info": analysis_content_parsed.get('data_info', {}),
"agents_analysis": analysis_content_parsed.get('agents_analysis', {}),
"scoring_ranking": pd.DataFrame(analysis_content_parsed.get('scoring_ranking', [])) if analysis_content_parsed.get('scoring_ranking') else None,
"final_report": analysis_content_parsed.get('final_report', {}),
"recommended_stocks": report_detail.get('recommended_stocks', [])
}
st.session_state.longhubang_result = loaded_result
st.success('✅ 报告已加载到分析页面，请切换到"龙虎榜分析"标签查看')
except Exception as e:
st.error(f"❌ 加载历史报告失败: {str(e)}")
import traceback
st.code(traceback.format_exc())
def display_statistics_tab():
@@ -790,6 +1013,365 @@ def display_statistics_tab():
st.error(f"❌ 加载统计数据失败: {str(e)}")
def run_longhubang_batch_analysis():
"""执行龙虎榜TOP股票批量分析（遵循统一调用规范）"""
st.markdown("## 🚀 龙虎榜TOP股票批量分析")
st.markdown("---")
# 检查是否已有分析结果
if st.session_state.get('longhubang_batch_results'):
display_longhubang_batch_results(st.session_state.longhubang_batch_results)
# 返回按钮
col_back, col_clear = st.columns(2)
with col_back:
if st.button("🔙 返回龙虎榜分析", use_container_width=True):
# 清除所有批量分析相关状态
if 'longhubang_batch_trigger' in st.session_state:
del st.session_state.longhubang_batch_trigger
if 'longhubang_batch_codes' in st.session_state:
del st.session_state.longhubang_batch_codes
if 'longhubang_batch_results' in st.session_state:
del st.session_state.longhubang_batch_results
st.rerun()
with col_clear:
if st.button("🔄 重新分析", use_container_width=True):
# 清除结果，保留触发标志和代码
if 'longhubang_batch_results' in st.session_state:
del st.session_state.longhubang_batch_results
st.rerun()
return
# 获取股票代码列表
stock_codes = st.session_state.get('longhubang_batch_codes', [])
if not stock_codes:
st.error("未找到股票代码列表")
# 清除触发标志
if 'longhubang_batch_trigger' in st.session_state:
del st.session_state.longhubang_batch_trigger
return
st.info(f"即将分析 {len(stock_codes)} 只股票：{', '.join(stock_codes)}")
# 返回按钮
if st.button("🔙 取消返回", type="secondary"):
# 清除所有批量分析相关状态
if 'longhubang_batch_trigger' in st.session_state:
del st.session_state.longhubang_batch_trigger
if 'longhubang_batch_codes' in st.session_state:
del st.session_state.longhubang_batch_codes
st.rerun()
st.markdown("---")
# 分析选项
col1, col2 = st.columns(2)
with col1:
analysis_mode = st.selectbox(
"分析模式",
options=["sequential", "parallel"],
format_func=lambda x: "顺序分析（稳定）" if x == "sequential" else "并行分析（快速）",
help="顺序分析较慢但稳定，并行分析更快但消耗更多资源"
)
with col2:
if analysis_mode == "parallel":
max_workers = st.number_input(
"并行线程数",
min_value=2,
max_value=5,
value=3,
help="同时分析的股票数量"
)
else:
max_workers = 1
st.markdown("---")
# 开始分析按钮
col_confirm, col_cancel = st.columns(2)
start_analysis = False
with col_confirm:
if st.button("🚀 确认开始分析", type="primary", use_container_width=True):
start_analysis = True
with col_cancel:
if st.button("❌ 取消", type="secondary", use_container_width=True):
# 清除所有批量分析相关状态
if 'longhubang_batch_trigger' in st.session_state:
del st.session_state.longhubang_batch_trigger
if 'longhubang_batch_codes' in st.session_state:
del st.session_state.longhubang_batch_codes
st.rerun()
if start_analysis:
# 导入统一分析函数（遵循统一规范）
from app import analyze_single_stock_for_batch
import concurrent.futures
import time
st.markdown("---")
st.info("⏳ 正在执行批量分析，请稍候...")
# 进度显示
progress_bar = st.progress(0)
status_text = st.empty()
results = []
start_time = time.time()
if analysis_mode == "sequential":
# 顺序分析
for i, code in enumerate(stock_codes):
status_text.text(f"正在分析 {code} ({i+1}/{len(stock_codes)})")
progress_bar.progress((i + 1) / len(stock_codes))
try:
# 调用统一分析函数
result = analyze_single_stock_for_batch(
symbol=code,
period="1y",
enabled_analysts_config={
'technical': True,
'fundamental': True,
'fund_flow': True,
'risk': True,
'sentiment': False,
'news': False
},
selected_model='deepseek-chat'
)
results.append({
"code": code,
"result": result
})
except Exception as e:
results.append({
"code": code,
"result": {"success": False, "error": str(e)}
})
else:
# 并行分析
status_text.text(f"并行分析 {len(stock_codes)} 只股票...")
def analyze_one(code):
try:
result = analyze_single_stock_for_batch(
symbol=code,
period="1y",
enabled_analysts_config={
'technical': True,
'fundamental': True,
'fund_flow': True,
'risk': True,
'sentiment': False,
'news': False
},
selected_model='deepseek-chat'
)
return {"code": code, "result": result}
except Exception as e:
return {"code": code, "result": {"success": False, "error": str(e)}}
with concurrent.futures.ThreadPoolExecutor(max_workers=max_workers) as executor:
futures = {executor.submit(analyze_one, code): code for code in stock_codes}
completed = 0
for future in concurrent.futures.as_completed(futures):
completed += 1
progress_bar.progress(completed / len(stock_codes))
status_text.text(f"已完成 {completed}/{len(stock_codes)}")
results.append(future.result())
# 清除进度
progress_bar.empty()
status_text.empty()
# 计算统计
elapsed_time = time.time() - start_time
success_count = sum(1 for r in results if r.get("result", {}).get("success"))
failed_count = len(results) - success_count
st.success(f"✅ 批量分析完成！成功 {success_count} 只，失败 {failed_count} 只，耗时 {elapsed_time:.1f}秒")
# 保存结果到session_state
st.session_state.longhubang_batch_results = {
"results": results,
"total": len(results),
"success": success_count,
"failed": failed_count,
"elapsed_time": elapsed_time
}
time.sleep(0.5)
st.rerun()
def display_longhubang_batch_results(batch_results: dict):
"""显示龙虎榜批量分析结果"""
st.markdown("### 📊 批量分析结果")
results = batch_results.get("results", [])
total = batch_results.get("total", 0)
success = batch_results.get("success", 0)
failed = batch_results.get("failed", 0)
elapsed_time = batch_results.get("elapsed_time", 0)
# 统计信息
col1, col2, col3, col4 = st.columns(4)
with col1:
st.metric("总计", total)
with col2:
st.metric("成功", success)
with col3:
st.metric("失败", failed)
with col4:
st.metric("耗时", f"{elapsed_time:.1f}秒")
st.markdown("---")
# 失败的股票
failed_results = [r for r in results if not r.get("result", {}).get("success")]
if failed_results:
with st.expander(f"❌ 失败股票 ({len(failed_results)}只)", expanded=False):
for item in failed_results:
code = item.get("code", "")
error = item.get("result", {}).get("error", "未知错误")
st.error(f"**{code}**: {error}")
# 成功的股票
success_results = [r for r in results if r.get("result", {}).get("success")]
if not success_results:
st.warning("⚠️ 没有成功分析的股票")
return
st.markdown("### 🎯 分析结果详情")
# 显示每只股票的分析结果（使用统一字段名）
for item in success_results:
code = item.get("code", "")
result = item.get("result", {})
final_decision = result.get("final_decision", {})
stock_info = result.get("stock_info", {})
# 使用统一字段名
rating = final_decision.get("rating", "未知")
confidence = final_decision.get("confidence_level", "N/A")
entry_range = final_decision.get("entry_range", "N/A")
take_profit = final_decision.get("take_profit", "N/A")
stop_loss = final_decision.get("stop_loss", "N/A")
target_price = final_decision.get("target_price", "N/A")
advice = final_decision.get("advice", "")
# 评级颜色
if "强烈买入" in rating or "买入" in rating:
rating_color = "🟢"
elif "卖出" in rating:
rating_color = "🔴"
else:
rating_color = "🟡"
with st.expander(f"{rating_color} {code} {stock_info.get('name', '')} - {rating} (信心度: {confidence})", expanded=False):
col1, col2, col3 = st.columns(3)
with col1:
st.markdown("**基本信息**")
st.write(f"当前价: {stock_info.get('current_price', 'N/A')}")
st.write(f"目标价: {target_price}")
with col2:
st.markdown("**进出场位置**")
st.write(f"进场区间: {entry_range}")
st.write(f"止盈位: {take_profit}")
with col3:
st.markdown("**风控**")
st.write(f"止损位: {stop_loss}")
st.write(f"评级: {rating}")
if advice:
st.markdown("**投资建议**")
st.info(advice)
# 添加到监测按钮
if st.button(f"➕ 加入监测", key=f"add_monitor_{code}"):
add_to_monitor_from_longhubang(code, stock_info.get('name', ''), final_decision)
def add_to_monitor_from_longhubang(code: str, name: str, final_decision: dict):
"""从龙虎榜分析结果添加到监测列表"""
try:
from monitor_db import monitor_db
import re
# 提取数据（使用统一字段名和解析逻辑）
rating = final_decision.get("rating", "持有")
entry_range = final_decision.get("entry_range", "")
take_profit_str = final_decision.get("take_profit", "")
stop_loss_str = final_decision.get("stop_loss", "")
# 解析进场区间
entry_min, entry_max = None, None
if entry_range and isinstance(entry_range, str) and "-" in entry_range:
try:
parts = entry_range.split("-")
entry_min = float(parts[0].strip())
entry_max = float(parts[1].strip())
except:
pass
# 解析止盈止损
take_profit, stop_loss = None, None
if take_profit_str:
try:
numbers = re.findall(r'\d+\.?\d*', str(take_profit_str))
if numbers:
take_profit = float(numbers[0])
except:
pass
if stop_loss_str:
try:
numbers = re.findall(r'\d+\.?\d*', str(stop_loss_str))
if numbers:
stop_loss = float(numbers[0])
except:
pass
# 验证必需参数
if not all([entry_min, entry_max, take_profit, stop_loss]):
st.error("❌ 分析结果缺少完整的进场区间和止盈止损信息")
return
# 添加到监测
monitor_db.add_monitored_stock(
symbol=code,
name=name,
rating=rating,
entry_range={"min": entry_min, "max": entry_max},
take_profit=take_profit,
stop_loss=stop_loss,
check_interval=60,
notification_enabled=True
)
st.success(f"✅ {code} 已成功加入监测列表！")
except Exception as e:
st.error(f"❌ 添加监测失败: {str(e)}")
# 测试函数
if __name__ == "__main__":
st.set_page_config(